1. Field
The present disclosure generally relates to detecting automated identities in computer user communities.
2. Description of the Related Art
Social networks are often plagued with people setting up fake identities to gather social interactions and other user data. This data can be gathered for various nefarious purposes such as spamming and identity theft.
One prevalent type of fake identity is a “robot” or automated identity that runs at least partially autonomously on a computing device.
Various techniques have been tried to weed out automated identities. Some of these techniques include asking a user to enter text matching noisy or distorted displayed characters. The concept behind this approach is that computer programs can have a difficult time using optical character recognition (OCR) techniques to read the noisy or distorted characters, whereas people can usually read the characters fairly easily. However, as OCR techniques have improved, the effectiveness of this type of approach has decreased.